PROBLEM WITH ANTHROPOMETRIC MEASURES
Assessing Child Nutrition: Problems with Anthropometric
Measures as a Proxy for Child Health in Malnourished children in imo state
CHAPTER 1
1.0 INTRODUCTION
1.1 OBJECTIVE OF THE STUDY
To improve the health condition of children in Imo State
To educate the children on the effect of under and over nutrition.
To ascertain the academic performance of students
1.2 SIGNIFICANT OF STUDY
1 This study will help the student, lecturers, parents/ guardians, government and others to identify the significant of nutrient requirements, such as vitamins and minerals, for the optimum growth of health status of children in undernourished population
2 This study will contribute towards the academic performance of the students and enhance the students effectiveness toward the right choice of food selection that contains the basic micro nutrients.
3 It will also help the policy planners in getting nutritional intervention programme for students.
4 It also help nutrition education to know where and how to channel his/her education programme
CHAPTER TWO
2.0 LITERATURE REVIEW
Assessing the health of children in developing countries is an imperative goal of national and international organizations, both because of the need to appropriately structure interventions (Becker and Black 1996; United Nations 1998), and because the health of children is a mirror on the health, social standing, and economic resources of their parents (DeRose, Das and Millman 2000; Kahn 2002; Winikoff 1988).
Studies have established the relationship between malnutrition and death in Matlab (Bhuiya et al. 1989), and poor nutrition in the past was a leading cause of increased overall mortality (Fauveau and Briend in Fauveau 1994). However, declines in child mortality in the region mean that measures of morbidity are more appropriate to understand the health of the population. As morbidity declined, differentials in health were documented and attributed to differential treatment of children in the home and selective access to health care, which are controlled by parents (Amin 1990, Chen, Huq, & D’Souza 1981, Coale & Banister 1994, Hill & Upchurch 1995, Johansson & Nygren 1991, Muhuri & Menken 1997, Muhuri and Preston 1991). While health disadvantage for girls has been documented based on gender preferences for sons and low maternal autonomy in the home, the results of recent studies are mixed as to whether selective neglect remains problematic, or whether sex bias favoring boys largely is fading.
CHAPTER THREE
3.0 MATERIALS AND METHODS
3.1 STUDY AREA:
The population under study comes from imo state, a society where environmental hazards are high and community health and educational infrastructure is poorly developed. The overwhelming majority of older individuals live with adult children (mostly sons) and alternative sources of support--financial and otherwise--outside the family are scarce. The predominant occupations for rural males are agricultural, with labor force participation rates remaining very high even for older males. Women are largely restricted by convention to activities within the home with relatively little opportunity to venture outside the homestead, although these restrictions are diminishing. Given the high level of poverty and the scarcity of health providers, contact with the formal health care system is thought to be relatively infrequent. It is a multistage,
3.2 SAMPLE SELECTION:
The sample that i use collected information from over 11,000 individuals aged 15 and over and nearly 5,000 children aged 1-14 in 4,538 households.
3.2.1 SURVEY DESIGN:
Random sampling were used in the selection of the target group, from imo state, owerri. up to two households were selected for detailed interviews. Within each selected household, two children were selected at random for further data collection.
3.2.2 SAMPLE SIZE DETERMINATION:
In determining the sample size to be used in the study work, I applied the formula of sample size determination in other to get maximum accuracy of the study work, the formula is shown below:
n = to determine the number of people that I want to work with.
n= sample size
N= population size
1 = constant
℮ = margin of error test of significance.
I assume that the total population in faculty of agriculture is and the level of significance is 0.05
n = to determine the number of people that I want to work with.
n= sample size
N= population size
1 = constant
℮ = margin of error test of significance.
I assume that the total population in imo state university is 11000 and the level of significance is 0.05
n =
n =
n = 392.2
n≈ 392
Therefore, I decided to work with 5000 students.
3.3 VALIDATION OF METHOD OF STUDY
A questionnaire was personally administered to the children in the sampled community in form of face to face, and information about their dietary intake, were collected. For communities with two or fewer households, all households were chosen. For baris with more than two households, the first household was chosen at random; the second household was selected from the bari in order of preference as follows: (i) the household of the father and/or mother of the head of the first sampled household, (ii) a household containing a son of the head of the first sampled household (chosen at random if there are multiple sons in separate households in the bari), (iii) a household containing a brother of the head of the first sampled household (chosen at random if there are multiple brothers in separate households in the bari) and (iv) a second randomly selected household. Probability weights are available for each individual included in the bari sample. The validated questionnaire was pretest to the children in imo state, owerri to ensure its reliability.Therefore, I decided to work with 5000 children
3.3.1 LIMITATION
This study lasted for two weeks
3.4 DATA COLLECTION
A structured and validated questionnaire was used to collect data. The children were visited in their homes and data was collected through the use of questionnaire. The questionnaire was designed to elicit information on the Assessing Child Nutrition: Problems with Anthropometric Measures as a Proxy for Child Health in Malnourished children in imo state.Responses on the questionnaire were analyse using computer package programme statistical package for social sciences (spss). Frequency distribution and percentage of the of the variables were calculated and tabulated.
CHAPTER FOUR
4.0 RESULT/FINDINGS
Malnourishment—low-weight-for-age—based on the current WHO standard is more prevalent in this population than stunting—low-height-for-age. The proportion of children who are stunted increases with age, and most stunted children also are malnourished (see Table 3). Girls and boys aged 1-4 experience good nutrition (normal on both height- and weight-for-age) in roughly equal proportions, confirming earlier work documenting disappearing sex-bias in nutrition (Trapp et. al 2004). At older ages, somewhat fewer girls are in the normal range and slightly higher proportions of girls have low weight-for-age. (Table 3 about here) Whereas young children experience the highest rates of acute sickness, chronic sickness increases with age, at least for boys (see Table 4). Table 4 gives the proportion of children with acute and chronic sickness by nutritional status. The rates of illness are quite high. Interestingly, children who are stunted (either low height-for-age only or both low height- and weight-for age)have lower rates of acute illness. (Tables 4 & 5 about here) Malnutrition does not correspond with reported illness as expected (see Table 5). In fact, children who are stunted, whether or not they also have low weight-for-age, actually have lower rates of acute sickness than those who have normal weight-for-age. It is possible that this finding is related to parental education or economic status differences in reporting; it is, however, more likely that it is an effect of age, in that older children have much higher rates of stunting and also lower rates of acute illness (as seen in Tables 3 and 4). In children below ten years old, low weight-for-age and height-for-age are related only to chronic illness, a relationship that does not hold for older children. Further, no relationship is found between illness of any type and low BMI.
Acute Sickness
BMI proves a poor measure of health in this population, which is undernourished by United States standards. Table 6 demonstrates that, because such a high proportion of children have both low weight- and height-for-age (68%), only just over a one third of these children are identified as having low BMI-for-age. Therefore, children with “normal” BMI-for-age may be thought of as composed of two subgroups: those with both normal height-for-age and weight-forage and those who are low on both in a way that balances out, so that their BMI is not unusually low. (Table 6 about here) None of the three measures predicted acute illness in any significant manner. Because the results are similar for BMI, weight-for-age, and height-for-age, we show only those for BMI in Table 7. Model 1 has controls for age and sex, and age operates in the expected direction with a coefficient of -0.08, which translates into an odds ratio of .917, indicating that the odds of acute illness decline 8% for each year of age. We next add the number of children in the household to the model, followed by parental education, household income, and MCH-FP program area residence. None of these controls improved the predictive power of the anthropometric measures, however age and MCH-FP area residence operate in the expected direction in relation to acute illness, providing some confirmation of the validity of the morbidity measures. However, in the final model, the odds of acute illness for those with low BMI differs from those without low BMI by only 13% (based on a coefficient of -135.) If BMI truly were an accurate predictor of poor health, we might expect an odds ratio of 60% or higher. (Note: Identical models were run using low weight-, height-for-age as a dependent variable, with identical relationships to those displayed for BMI). In addition to the models shown, we tested various family structure characteristics listed in Table 5, none of which were significant.
Chronic Illness
Similar analyses attempted to predict chronic illness from the three proxy health measures. Interestingly, only low weight-for-age predicts chronic illness, rather than low height- or BMI-for-age as expected (these models are not shown, but demonstrate no significant relationships). Contrary to expectations, underweight appears to predict chronic illnesses in this population, a relationship that persists with controls for age, sex, parental education, income, and MCH-FP program area residence, none of which are significant (see Table 8). However, because of the strong increase in low weight, height, and BMI with age, we did not end the investigation with these models. (Table 8 about here) We next tested the relationship between health proxies and health separately for children over and under the age of 10. When children under age 10 are considered separate from adolescents, both low height-for-age and low weight-for-age are significant predictors of increased chronic illness (see Table 9). BMI still is not a predictor of chronic illness for young children. Nor are any of the three measures predictive for children 10-14. (Table 9 about here) Because the results shown in Table 9 indicate that weight- or height-for-age are predictors of chronic illness in younger children while BMI was not, we examined the relationship further. As mentioned earlier, BMI can be in the normal range in two ways: both weight and height are in the normal range for age or both weight and height are low for age but, balancing one another, lead to normal BMI for age. This is borne out in Table 10, which shows that, of the children in this study who have normal BMI (<2SD below the median-for-age), over 40% are both stunted and wasted, as compared to 48% of the children who are assessed as having low BMI. (Table 10 about here) Table 11 shows models in which first height-for-age or weight-for-age was included, along with age and sex for controls (Models 1 and 3). Then an indicator of whether a child was both low weight and low height was added (Models 2 and 4). In each model, only the latter variable was significant predictor. Finally, Model 5 shows that the model in which only “both” is entered fits as well as the models with either low height or low weight added.
Table 1: Proportion of wasted and stunted children by age in 1978 and 1996 (Using Previous
Standard)
CHAPTER 5
5.0 DISCUSSION
This analysis contributes to our understanding of child health in developing countries in three significant ways. First, we demonstrate that in undernourished populations, BMI is not necessarily a reliable measure of poor health in children. Under those circumstances, children whose BMI is in the normal range may be both stunted and malnourished as measured by low weight-for age. In the current study, BMI predicts neither acute nor chronic illness in children under 15. For children under 10, it appears that those who are both low weight- and low height for age have the highest rates of chronic illness. These are the very measures that go into BMI, but combining them as a ratio removes their predictive power. Thus, whereas we find support for the work of Pande (2003) and others who find waning sex-bias in children’s health based on anthropometry, we remain concerned that in Bangladesh these “objective” measures fail to capture a child’s underlying nutritional status. Further, caution must be used because these proxies fail to correspond with poor health in some cases, they also fail to predict the expected relationship in terms of current or past morbidity. Second, this analysis demonstrates clearly the need for better health measures in adolescents, for whom puberty and biological differences in growth by sex confound easy assessment. We note that as children age in this generally malnourished population, they diverge from the reference population to such a degree that the usual cutoff for ill effects fails to provide sufficient distinction between children who suffer the effects of malnutrition, and those who merely are small and light.
5.1 CONCLUSIONS
The implications of these findings speak to how health is measured in undernourished populations such as those in imo state. First, anthropometric measures that are used as a proxy for health may be an unreliable yardstick for child wellness. While collecting actual data on food intake and allocation in the household can be difficult and expensive, and retrospective reports of morbidity can be unreliable, such measures may be necessary to more accurately gauge the health of populations of children, particularly in settings like Bangladesh where common measures of poor health (such as BMI) may not produce expected results. Biomarkers or other objective measures of health may be essential. The problem is complicated for adolescents, whose growth rates are known to diverge more widely than younger children, and for whom biological differences in growth rates by sex may confound any health proxy. Third, the reference population may provide a poor comparison by failing to provide adequate distinction between ill children and those who merely are small.
5.2 RECOMMENDATION
This study recommended that a better threshold would allow more accurate identification of children at risk using proxy measures, however, until a better reference is found and tested, proxies appear to be an unreliable tool. Given the link between malnourishment and death cited earlier (Bhuiya et al. 1989; Fauveau and Briend in Fauveau 1994), it is crucial to find a means of accurately separating children who are at risk of health problems from those who are small but otherwise healthy. Measurement problems continue to confound our understanding of child morbidity in developing countries, particularly those older than nine years.
5.3 REFERENCES
Amin, S. 1990. “The Effect of Women’s Status on Sex Differentials in Infant and Child Mortality in South Asia.” Genus 46:55-69.
Aziz, K. 1979. "Kinship in Bangladesh." Dacca: International Centre for Diarrhoeal Disease Research.
Bairagi, R. 1986. “Food Crisis, Nutrition, and Female Children in Rural Bangladesh.”Population and Development Review. 12:307.
Becker, S. & R. Black. 1996. “A Model of Child Morbidity, Mortality, and HealthInterventions.” Population and Development Review. 22: 431.
Behrman, J., A. Foster, M. Rosenzweig & P. Vashishtha. 1996. “Women’s Schooling, Home Teaching, and Economic Growth.” The Journal of Political Economy 107: 682.
Bhuiya, A., B. Wohtyniak, & R. Karim. 1989, “Mother’s Education and Survival of Female Children in a Rural Area of Bangladesh.” Population and Development Review. 7:55-70.
Chen, L., E. Huq, & S. D’Souza. 1981. “Sex Bias in the Family Allocation of Food and Health Care in Rural Bangladesh.” Population and Development Review. 7:55.
Cole, T., M. Bellizzi, K. Flegal, & W. Dietz. 2000. “Establishing a Standard Definition for Child Overweight and Obesity Worldwide.” British Medical Journal 320:1-6.
D’Souza S. & L. Chen. 1980. “Sex Differentials in Mortality in Rural Bangladesh.” Population and Development Review 6:257-270.
DeRose, L., M. Das & S. Millman. 2000 “Does Female Disadvantage Mean Lower Access to Food?” Population and Development Review 26: 517-547.
Desai, S. 1992. “Children at Risk: The Role of Family Structure in Latin America and West Africa.” Population and Development Review 689:717.Fauveau, V. 1994. Matlab: Women, Children and Health. ICDDR Special Publication. July. 35.
Hill, K. & D. Upchurch. 1995. “Gender Differences in Child Health: Evidence from the Demographic and Health Surveys.” Population and Development Review 21:127-151.
Kahn, N. 2002. Health Care and Family Support for the Elderly in Aging Developing Countries:
An International Review and a Case Study of Health Care Allocation in Rural Bangladesh. Unpublished dissertation.
Kuhn, R., O. Rahman, & J. Menken. 2004. “Relating Self-Reported and Objective HealthIndicators to Adult Mortality in Bangladesh.” Population Aging Center Working Paper
2004-0004. University of Colorado Institute of Behavioral Science.Legrand, T. & J. Phillips. 1996. “The Effect of Fertility Reduction on Infant and Child Mortality:
Evidence from Matlab in Rural Bangladesh.” Population Studies. 50:51.Madise, N., Z. Matthews, & B. Margetts. 1999. “Heterogeneity of Child Nutritional Status Between Households: A Comparison of Six Sub-Saharan African Countries.” Population Studies 53:331-343.
Menken, J., & J. Phillips. 1990. “Population Change in a Rural Area of Bangladesh, 1967-87”.
Annals of the American Academy of Political and Social Science 510: 87-101.
Menken, J., L. Duffy, & R. Kuhn. 2003. “Childbearing and Women’s Survival in rural Bangladesh.” Population and Development Review 29: 405-426.
Menken, J. & E. Trapp. 2004.”Differential Treatment of Children by Sex.” (forthcoming) In Encyclopedia of Sociology. George Ritzer (ed.)
Mosley, W. & L. Chen. 1984. “An Analytical Framework for the Study of Child Survival in Developing Countries.” Population and Development Review 10S: 25-45.
Mostafa, G., M.A.K. Shaikh, J. van Ginneken & A. M. Sardar. 1998. “Demographic Surveillance System—Matlab. Scientific Report No. 82.” ICDDR,B: Dhaka, Bangladesh.
Muhuri, P. 1995. “Health Programs, Maternal Education, and Differential Child Mortality in Matlab, Bangladesh.” Population and Development Review. 21: 813-834.
Muhuri, P. & J. Menken. 1997. “Adverse Effects of Next Birth, Gender, and FamilyComposition on Child Survival in Rural Bangladesh.” Population Studies. 51:279.
Muhuri, P. & S. Preston. 1991. “Effects of Family Composition on Mortality Differentials by Sex Among Children in Matlab, Bangladesh.” Population and Development Review 17:415-434.
Pande, R. 2003. “Selective Gender Differences in Childhood Nutrition and Immunization in Rural India: The Role of Siblings.” Demography 40: 395-418.
Pelletier, D. 1998. “Malnutrition, Morbidity and Child Mortality in Developing Countries.” In Too Young to Die: Genes or Gender? New York: United Nations.
Pelletier, D, E. Frongillo Jr., J. Habicht. 1993. “Epidemiologic Evidence for a Potentiating Effect of Malnutrition on Child Mortality.” American Journal of Public Health 83:1130-1133.
Rahman, M.O., J. Menken, A. Foster, C. Peterson, M.N. Khan, R. Kuhn, & P. Gertler. 1999.The Matlab Health and Socioeconomic Survey: Overview and User’s Guide, DRU-2018/1. Santa Monica: RAND.
Shaikh, K., G. Mostafa, A. Bhuiya, A. M. Sardar, I. Molla, & B. Wojtyniak. 1985. “Demographic Surveillance System—Matlab. Scientific Report No. 64.” ICDDR,B: Dhaka, Bangladesh.
Thomas, D., E. Frankenberg, K. Beegle, & G. Teruel. 1999. Household budgets, householdconsumption and the crisis in Indonesia: Evidence from longitudinal household surveydata. Los Angeles, CA: RAND. United Nations. 1998. Too Young to Die: Genes or Gender? New York: United Nations
van Ginneken, J., R. Bairagi, A. Francisco, A Sarder, & P. Vaughan. 1998. Health and Demographic Surveillance in Matlab: Past, Present and Future. Dhaka: ICDDR.
Waterlow, J.C. 1972. “Classification and Definition of Protein-Calorie Malnutrition.” British Medical Journal. 3: 566.
Winikoff, B. 1988. “Women’s Health: An Alternative Perspective for Choosing Interventions.” Studies in Family Planning 19:197-214.
World Health Organization. 1997. “Who Global Database on Child Growth and Malnutrition.” Document WHO/NUT/97.4.
Measures as a Proxy for Child Health in Malnourished children in imo state
CHAPTER 1
1.0 INTRODUCTION
1.1 OBJECTIVE OF THE STUDY
To improve the health condition of children in Imo State
To educate the children on the effect of under and over nutrition.
To ascertain the academic performance of students
1.2 SIGNIFICANT OF STUDY
1 This study will help the student, lecturers, parents/ guardians, government and others to identify the significant of nutrient requirements, such as vitamins and minerals, for the optimum growth of health status of children in undernourished population
2 This study will contribute towards the academic performance of the students and enhance the students effectiveness toward the right choice of food selection that contains the basic micro nutrients.
3 It will also help the policy planners in getting nutritional intervention programme for students.
4 It also help nutrition education to know where and how to channel his/her education programme
CHAPTER TWO
2.0 LITERATURE REVIEW
Assessing the health of children in developing countries is an imperative goal of national and international organizations, both because of the need to appropriately structure interventions (Becker and Black 1996; United Nations 1998), and because the health of children is a mirror on the health, social standing, and economic resources of their parents (DeRose, Das and Millman 2000; Kahn 2002; Winikoff 1988).
Studies have established the relationship between malnutrition and death in Matlab (Bhuiya et al. 1989), and poor nutrition in the past was a leading cause of increased overall mortality (Fauveau and Briend in Fauveau 1994). However, declines in child mortality in the region mean that measures of morbidity are more appropriate to understand the health of the population. As morbidity declined, differentials in health were documented and attributed to differential treatment of children in the home and selective access to health care, which are controlled by parents (Amin 1990, Chen, Huq, & D’Souza 1981, Coale & Banister 1994, Hill & Upchurch 1995, Johansson & Nygren 1991, Muhuri & Menken 1997, Muhuri and Preston 1991). While health disadvantage for girls has been documented based on gender preferences for sons and low maternal autonomy in the home, the results of recent studies are mixed as to whether selective neglect remains problematic, or whether sex bias favoring boys largely is fading.
CHAPTER THREE
3.0 MATERIALS AND METHODS
3.1 STUDY AREA:
The population under study comes from imo state, a society where environmental hazards are high and community health and educational infrastructure is poorly developed. The overwhelming majority of older individuals live with adult children (mostly sons) and alternative sources of support--financial and otherwise--outside the family are scarce. The predominant occupations for rural males are agricultural, with labor force participation rates remaining very high even for older males. Women are largely restricted by convention to activities within the home with relatively little opportunity to venture outside the homestead, although these restrictions are diminishing. Given the high level of poverty and the scarcity of health providers, contact with the formal health care system is thought to be relatively infrequent. It is a multistage,
3.2 SAMPLE SELECTION:
The sample that i use collected information from over 11,000 individuals aged 15 and over and nearly 5,000 children aged 1-14 in 4,538 households.
3.2.1 SURVEY DESIGN:
Random sampling were used in the selection of the target group, from imo state, owerri. up to two households were selected for detailed interviews. Within each selected household, two children were selected at random for further data collection.
3.2.2 SAMPLE SIZE DETERMINATION:
In determining the sample size to be used in the study work, I applied the formula of sample size determination in other to get maximum accuracy of the study work, the formula is shown below:
n = to determine the number of people that I want to work with.
n= sample size
N= population size
1 = constant
℮ = margin of error test of significance.
I assume that the total population in faculty of agriculture is and the level of significance is 0.05
n = to determine the number of people that I want to work with.
n= sample size
N= population size
1 = constant
℮ = margin of error test of significance.
I assume that the total population in imo state university is 11000 and the level of significance is 0.05
n =
n =
n = 392.2
n≈ 392
Therefore, I decided to work with 5000 students.
3.3 VALIDATION OF METHOD OF STUDY
A questionnaire was personally administered to the children in the sampled community in form of face to face, and information about their dietary intake, were collected. For communities with two or fewer households, all households were chosen. For baris with more than two households, the first household was chosen at random; the second household was selected from the bari in order of preference as follows: (i) the household of the father and/or mother of the head of the first sampled household, (ii) a household containing a son of the head of the first sampled household (chosen at random if there are multiple sons in separate households in the bari), (iii) a household containing a brother of the head of the first sampled household (chosen at random if there are multiple brothers in separate households in the bari) and (iv) a second randomly selected household. Probability weights are available for each individual included in the bari sample. The validated questionnaire was pretest to the children in imo state, owerri to ensure its reliability.Therefore, I decided to work with 5000 children
3.3.1 LIMITATION
This study lasted for two weeks
3.4 DATA COLLECTION
A structured and validated questionnaire was used to collect data. The children were visited in their homes and data was collected through the use of questionnaire. The questionnaire was designed to elicit information on the Assessing Child Nutrition: Problems with Anthropometric Measures as a Proxy for Child Health in Malnourished children in imo state.Responses on the questionnaire were analyse using computer package programme statistical package for social sciences (spss). Frequency distribution and percentage of the of the variables were calculated and tabulated.
CHAPTER FOUR
4.0 RESULT/FINDINGS
Malnourishment—low-weight-for-age—based on the current WHO standard is more prevalent in this population than stunting—low-height-for-age. The proportion of children who are stunted increases with age, and most stunted children also are malnourished (see Table 3). Girls and boys aged 1-4 experience good nutrition (normal on both height- and weight-for-age) in roughly equal proportions, confirming earlier work documenting disappearing sex-bias in nutrition (Trapp et. al 2004). At older ages, somewhat fewer girls are in the normal range and slightly higher proportions of girls have low weight-for-age. (Table 3 about here) Whereas young children experience the highest rates of acute sickness, chronic sickness increases with age, at least for boys (see Table 4). Table 4 gives the proportion of children with acute and chronic sickness by nutritional status. The rates of illness are quite high. Interestingly, children who are stunted (either low height-for-age only or both low height- and weight-for age)have lower rates of acute illness. (Tables 4 & 5 about here) Malnutrition does not correspond with reported illness as expected (see Table 5). In fact, children who are stunted, whether or not they also have low weight-for-age, actually have lower rates of acute sickness than those who have normal weight-for-age. It is possible that this finding is related to parental education or economic status differences in reporting; it is, however, more likely that it is an effect of age, in that older children have much higher rates of stunting and also lower rates of acute illness (as seen in Tables 3 and 4). In children below ten years old, low weight-for-age and height-for-age are related only to chronic illness, a relationship that does not hold for older children. Further, no relationship is found between illness of any type and low BMI.
Acute Sickness
BMI proves a poor measure of health in this population, which is undernourished by United States standards. Table 6 demonstrates that, because such a high proportion of children have both low weight- and height-for-age (68%), only just over a one third of these children are identified as having low BMI-for-age. Therefore, children with “normal” BMI-for-age may be thought of as composed of two subgroups: those with both normal height-for-age and weight-forage and those who are low on both in a way that balances out, so that their BMI is not unusually low. (Table 6 about here) None of the three measures predicted acute illness in any significant manner. Because the results are similar for BMI, weight-for-age, and height-for-age, we show only those for BMI in Table 7. Model 1 has controls for age and sex, and age operates in the expected direction with a coefficient of -0.08, which translates into an odds ratio of .917, indicating that the odds of acute illness decline 8% for each year of age. We next add the number of children in the household to the model, followed by parental education, household income, and MCH-FP program area residence. None of these controls improved the predictive power of the anthropometric measures, however age and MCH-FP area residence operate in the expected direction in relation to acute illness, providing some confirmation of the validity of the morbidity measures. However, in the final model, the odds of acute illness for those with low BMI differs from those without low BMI by only 13% (based on a coefficient of -135.) If BMI truly were an accurate predictor of poor health, we might expect an odds ratio of 60% or higher. (Note: Identical models were run using low weight-, height-for-age as a dependent variable, with identical relationships to those displayed for BMI). In addition to the models shown, we tested various family structure characteristics listed in Table 5, none of which were significant.
Chronic Illness
Similar analyses attempted to predict chronic illness from the three proxy health measures. Interestingly, only low weight-for-age predicts chronic illness, rather than low height- or BMI-for-age as expected (these models are not shown, but demonstrate no significant relationships). Contrary to expectations, underweight appears to predict chronic illnesses in this population, a relationship that persists with controls for age, sex, parental education, income, and MCH-FP program area residence, none of which are significant (see Table 8). However, because of the strong increase in low weight, height, and BMI with age, we did not end the investigation with these models. (Table 8 about here) We next tested the relationship between health proxies and health separately for children over and under the age of 10. When children under age 10 are considered separate from adolescents, both low height-for-age and low weight-for-age are significant predictors of increased chronic illness (see Table 9). BMI still is not a predictor of chronic illness for young children. Nor are any of the three measures predictive for children 10-14. (Table 9 about here) Because the results shown in Table 9 indicate that weight- or height-for-age are predictors of chronic illness in younger children while BMI was not, we examined the relationship further. As mentioned earlier, BMI can be in the normal range in two ways: both weight and height are in the normal range for age or both weight and height are low for age but, balancing one another, lead to normal BMI for age. This is borne out in Table 10, which shows that, of the children in this study who have normal BMI (<2SD below the median-for-age), over 40% are both stunted and wasted, as compared to 48% of the children who are assessed as having low BMI. (Table 10 about here) Table 11 shows models in which first height-for-age or weight-for-age was included, along with age and sex for controls (Models 1 and 3). Then an indicator of whether a child was both low weight and low height was added (Models 2 and 4). In each model, only the latter variable was significant predictor. Finally, Model 5 shows that the model in which only “both” is entered fits as well as the models with either low height or low weight added.
Table 1: Proportion of wasted and stunted children by age in 1978 and 1996 (Using Previous
Standard)
CHAPTER 5
5.0 DISCUSSION
This analysis contributes to our understanding of child health in developing countries in three significant ways. First, we demonstrate that in undernourished populations, BMI is not necessarily a reliable measure of poor health in children. Under those circumstances, children whose BMI is in the normal range may be both stunted and malnourished as measured by low weight-for age. In the current study, BMI predicts neither acute nor chronic illness in children under 15. For children under 10, it appears that those who are both low weight- and low height for age have the highest rates of chronic illness. These are the very measures that go into BMI, but combining them as a ratio removes their predictive power. Thus, whereas we find support for the work of Pande (2003) and others who find waning sex-bias in children’s health based on anthropometry, we remain concerned that in Bangladesh these “objective” measures fail to capture a child’s underlying nutritional status. Further, caution must be used because these proxies fail to correspond with poor health in some cases, they also fail to predict the expected relationship in terms of current or past morbidity. Second, this analysis demonstrates clearly the need for better health measures in adolescents, for whom puberty and biological differences in growth by sex confound easy assessment. We note that as children age in this generally malnourished population, they diverge from the reference population to such a degree that the usual cutoff for ill effects fails to provide sufficient distinction between children who suffer the effects of malnutrition, and those who merely are small and light.
5.1 CONCLUSIONS
The implications of these findings speak to how health is measured in undernourished populations such as those in imo state. First, anthropometric measures that are used as a proxy for health may be an unreliable yardstick for child wellness. While collecting actual data on food intake and allocation in the household can be difficult and expensive, and retrospective reports of morbidity can be unreliable, such measures may be necessary to more accurately gauge the health of populations of children, particularly in settings like Bangladesh where common measures of poor health (such as BMI) may not produce expected results. Biomarkers or other objective measures of health may be essential. The problem is complicated for adolescents, whose growth rates are known to diverge more widely than younger children, and for whom biological differences in growth rates by sex may confound any health proxy. Third, the reference population may provide a poor comparison by failing to provide adequate distinction between ill children and those who merely are small.
5.2 RECOMMENDATION
This study recommended that a better threshold would allow more accurate identification of children at risk using proxy measures, however, until a better reference is found and tested, proxies appear to be an unreliable tool. Given the link between malnourishment and death cited earlier (Bhuiya et al. 1989; Fauveau and Briend in Fauveau 1994), it is crucial to find a means of accurately separating children who are at risk of health problems from those who are small but otherwise healthy. Measurement problems continue to confound our understanding of child morbidity in developing countries, particularly those older than nine years.
5.3 REFERENCES
Amin, S. 1990. “The Effect of Women’s Status on Sex Differentials in Infant and Child Mortality in South Asia.” Genus 46:55-69.
Aziz, K. 1979. "Kinship in Bangladesh." Dacca: International Centre for Diarrhoeal Disease Research.
Bairagi, R. 1986. “Food Crisis, Nutrition, and Female Children in Rural Bangladesh.”Population and Development Review. 12:307.
Becker, S. & R. Black. 1996. “A Model of Child Morbidity, Mortality, and HealthInterventions.” Population and Development Review. 22: 431.
Behrman, J., A. Foster, M. Rosenzweig & P. Vashishtha. 1996. “Women’s Schooling, Home Teaching, and Economic Growth.” The Journal of Political Economy 107: 682.
Bhuiya, A., B. Wohtyniak, & R. Karim. 1989, “Mother’s Education and Survival of Female Children in a Rural Area of Bangladesh.” Population and Development Review. 7:55-70.
Chen, L., E. Huq, & S. D’Souza. 1981. “Sex Bias in the Family Allocation of Food and Health Care in Rural Bangladesh.” Population and Development Review. 7:55.
Cole, T., M. Bellizzi, K. Flegal, & W. Dietz. 2000. “Establishing a Standard Definition for Child Overweight and Obesity Worldwide.” British Medical Journal 320:1-6.
D’Souza S. & L. Chen. 1980. “Sex Differentials in Mortality in Rural Bangladesh.” Population and Development Review 6:257-270.
DeRose, L., M. Das & S. Millman. 2000 “Does Female Disadvantage Mean Lower Access to Food?” Population and Development Review 26: 517-547.
Desai, S. 1992. “Children at Risk: The Role of Family Structure in Latin America and West Africa.” Population and Development Review 689:717.Fauveau, V. 1994. Matlab: Women, Children and Health. ICDDR Special Publication. July. 35.
Hill, K. & D. Upchurch. 1995. “Gender Differences in Child Health: Evidence from the Demographic and Health Surveys.” Population and Development Review 21:127-151.
Kahn, N. 2002. Health Care and Family Support for the Elderly in Aging Developing Countries:
An International Review and a Case Study of Health Care Allocation in Rural Bangladesh. Unpublished dissertation.
Kuhn, R., O. Rahman, & J. Menken. 2004. “Relating Self-Reported and Objective HealthIndicators to Adult Mortality in Bangladesh.” Population Aging Center Working Paper
2004-0004. University of Colorado Institute of Behavioral Science.Legrand, T. & J. Phillips. 1996. “The Effect of Fertility Reduction on Infant and Child Mortality:
Evidence from Matlab in Rural Bangladesh.” Population Studies. 50:51.Madise, N., Z. Matthews, & B. Margetts. 1999. “Heterogeneity of Child Nutritional Status Between Households: A Comparison of Six Sub-Saharan African Countries.” Population Studies 53:331-343.
Menken, J., & J. Phillips. 1990. “Population Change in a Rural Area of Bangladesh, 1967-87”.
Annals of the American Academy of Political and Social Science 510: 87-101.
Menken, J., L. Duffy, & R. Kuhn. 2003. “Childbearing and Women’s Survival in rural Bangladesh.” Population and Development Review 29: 405-426.
Menken, J. & E. Trapp. 2004.”Differential Treatment of Children by Sex.” (forthcoming) In Encyclopedia of Sociology. George Ritzer (ed.)
Mosley, W. & L. Chen. 1984. “An Analytical Framework for the Study of Child Survival in Developing Countries.” Population and Development Review 10S: 25-45.
Mostafa, G., M.A.K. Shaikh, J. van Ginneken & A. M. Sardar. 1998. “Demographic Surveillance System—Matlab. Scientific Report No. 82.” ICDDR,B: Dhaka, Bangladesh.
Muhuri, P. 1995. “Health Programs, Maternal Education, and Differential Child Mortality in Matlab, Bangladesh.” Population and Development Review. 21: 813-834.
Muhuri, P. & J. Menken. 1997. “Adverse Effects of Next Birth, Gender, and FamilyComposition on Child Survival in Rural Bangladesh.” Population Studies. 51:279.
Muhuri, P. & S. Preston. 1991. “Effects of Family Composition on Mortality Differentials by Sex Among Children in Matlab, Bangladesh.” Population and Development Review 17:415-434.
Pande, R. 2003. “Selective Gender Differences in Childhood Nutrition and Immunization in Rural India: The Role of Siblings.” Demography 40: 395-418.
Pelletier, D. 1998. “Malnutrition, Morbidity and Child Mortality in Developing Countries.” In Too Young to Die: Genes or Gender? New York: United Nations.
Pelletier, D, E. Frongillo Jr., J. Habicht. 1993. “Epidemiologic Evidence for a Potentiating Effect of Malnutrition on Child Mortality.” American Journal of Public Health 83:1130-1133.
Rahman, M.O., J. Menken, A. Foster, C. Peterson, M.N. Khan, R. Kuhn, & P. Gertler. 1999.The Matlab Health and Socioeconomic Survey: Overview and User’s Guide, DRU-2018/1. Santa Monica: RAND.
Shaikh, K., G. Mostafa, A. Bhuiya, A. M. Sardar, I. Molla, & B. Wojtyniak. 1985. “Demographic Surveillance System—Matlab. Scientific Report No. 64.” ICDDR,B: Dhaka, Bangladesh.
Thomas, D., E. Frankenberg, K. Beegle, & G. Teruel. 1999. Household budgets, householdconsumption and the crisis in Indonesia: Evidence from longitudinal household surveydata. Los Angeles, CA: RAND. United Nations. 1998. Too Young to Die: Genes or Gender? New York: United Nations
van Ginneken, J., R. Bairagi, A. Francisco, A Sarder, & P. Vaughan. 1998. Health and Demographic Surveillance in Matlab: Past, Present and Future. Dhaka: ICDDR.
Waterlow, J.C. 1972. “Classification and Definition of Protein-Calorie Malnutrition.” British Medical Journal. 3: 566.
Winikoff, B. 1988. “Women’s Health: An Alternative Perspective for Choosing Interventions.” Studies in Family Planning 19:197-214.
World Health Organization. 1997. “Who Global Database on Child Growth and Malnutrition.” Document WHO/NUT/97.4.